In the age of physical tape, we knew what “recording” was. But might modern digital devices listen without recording? CC-licensed photo by Carbon Arc on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

A selection of 9 links for you. Reely hard. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

“ASOS hacked”: app users receive notifications sent by hackers • BBC News

Joe Tidy and Liv McMahon:

»

Asos says it is investigating “unauthorised activity” involving third-party platforms it uses after customers received a notification from its app sent by hackers.

Dozens of people told the BBC they received the strange “ASOS HACKED” message from the clothing and beauty store’s app on Tuesday morning – with some saying it left them “scared” to open the app.

The notification was addressed to the company’s data protection officer and IT teams in what cyber security experts said looked like a “brazen” extortion attempt.

Asos acknowledged the “unauthorised customer notification” on Tuesday afternoon, saying some “basic personal information” may have been accessed. In an email to customers on Tuesday night, the company apologised and urged customers not to engage with the notification. And it said the website and app are “operating as usual” promising customers they can “shop with confidence” while it investigates the incident.

The company has not as of yet informed the UK’s data watchdog, the Information Commission’s Office (ICO), about any breach. Exactly how many Asos customers received the notification on Tuesday remains unclear, but Google’s Play store says the ASOS app has been downloaded to android devices more than 10 million times.

The British retailer has a substantial global footprint – serving around 17 million customers each year across more than 150 markets.

Some Asos app users in Australia, France, Sweden and the Republic of Ireland had also received the notification, according to local reports on Tuesday.

Hackers seeking to pile pressure on potential victims by informing their customers is rare, as most extortions happen in private, so this incident may go down as a significant moment in cyber-attack history.

Shares in the company fell by around a tenth on Tuesday.

«

Asos is in hot water for not acknowledging the hack, but the biggest concern is what it implies about how badly the company has been hacked.
unique link to this extract

Viral Fallout and MW2 AI decompilation projects hit with DMCA • The Gamer

Nazmul Roosevelt:

»

DMCA takedowns for mods and fan overhauls have been happening for years, so it makes sense that decompilation projects, especially the ones assembled quickly and shared widely on social media, are now drawing the same kind of attention. These projects have exploded in number over the past week or so, but many rely heavily on generative AI. Even well‑known efforts like Ship of Harkinian use some form of genAI, despite beginning long before the current wave of vibe‑coded recreations.

The projects that went viral this week were built with far more aggressive AI assistance. One of the first was Fallout: New York. ChrisFirst shared a clip on Twitter showing a browser‑playable recreation of Fallout built with Claude Opus 5.5. It had quests, dialogue trees, VATS, a working Pip‑Boy, and an entire city generated without any texture or sound files. Every building, weapon, face, and ambient noise was “generated by code.” The post pulled in millions of views before ChrisFirst shared that he received a cease-and-desist from Bethesda and shut the project down.

The second project that’s since been taken down came from Maurice Heumann, who used generative AI to decompile Modern Warfare 2 on PC. He said he had seventeen agents working at the same time, reconstructing the game faster than any manual effort could. According to him, the game was already playable, most bugs were fixed, and he had even earned a nuke. Activision responded by removing the footage from his post and blocking access to the site hosting the code. Heumann later shared that his AI agents kept deleting his files.

The difference seems to come with scale. Ocarina of Time running in ultrawide at 280 frames is a transformative example of what was previously capable on native hardware. It’s impressive to see, even if the game was never meant to be viewed that way. The strongest decomp projects show evidence of careful manual work, but the rapid spread of AI‑assisted recreations is pushing the space into territory that publishers can no longer ignore.

«

What’s happening is that gamers can now use AI agents to decompile the executables of games, and then get those recompiled for other platforms, or add new elements that weren’t there before. And publishers don’t like it when such work is made public. (Obviously they can’t see any private work.)

Feels like this might drive games companies to cloud-based services which can’t be decompiled because they’re never running on the user’s computer.
unique link to this extract

Silicon Valley thinks AI will kill jobs. Economists are not convinced • Financial Times

Delphine Strauss:

»

The most worrying development is a cross-country slowdown in entry-level hiring in tech-related occupations — consistent with AI turbocharging the productivity of senior staff while taking over routine tasks done by juniors.

But there is no consensus on the cause. “Unemployment of young graduates has worsened since 2015. It’s a long-term trend. AI is clearly not the only factor,” Broecke argued, citing a potential oversupply of graduates, the rise of remote work and high interest rates.

The absence of aggregate job losses may simply be the calm before the storm. “Talk to executives in the non-tech sector — those who use AI as customers — and many say they don’t need as many workers as they did before AI,” Johnson said. “It is not yet in the numbers . . . but the automation threat is very real.”

Modelling by the Anthropic Institute, published last month, included an extreme scenario where rapid gains in AI’s capabilities cut “cognitive” employment by a fifth by 2030, leaving one in five white-collar workers and one in 10 of all workers unemployed.

The model is highly sensitive to underlying assumptions. The extreme scenario assumes AI will rapidly become able to perform most tasks without human oversight, will be widely adopted and will generate big productivity gains.

Other scenarios show more modest change, but Anton Korinek, who led the modelling, said he “personally put more weight” on those showing a rapid increase in AI adoption from 2028, with the technology becoming able to rival and replace senior executives, not just entry-level staff.

In practice, companies may be slow to shed workers even when AI tools can do the job. Broecke of the OECD said: “There is the cost of technology. There are legal barriers, social preferences, ethical issues. There are all sorts of reasons why something that is theoretically possible will not happen in practice.”

«

unique link to this extract

Google told to halt work on datacentres in Finland over environmental concerns • The Guardian

Miranda Bryant:

»

Google has been ordered to temporarily stop work on two datacentres in Finland after failing to do the necessary environmental impact assessments on its construction sites.

The planned datacentres in Muhos and Kajanni are part of a €13bn (£11bn) project across multiple sites in northern Finland by the technology company amid huge demand for space to house online infrastructure.

It comes after it emerged that Tuike Finland Oy, the company representing Google, had felled 330 hectares (815 acres) of forest in Muhos without completing an environmental impact assessment (EIA).

In Kajaani, the company has reportedly been logging across just under 200 hectares (494 acres) of land.

The Finnish Association for Nature Conservation said the sites had been reduced to “sand”, blaming the oversights on the “overheated” market for new datacentres.

On Tuesday, the Finnish Licensing and Supervision Authority (LVV) said it had issued a notice to the company to “immediately suspend” – by 23 October at the latest – “preparatory measures that significantly change the environment” of the two sites until the environmental impact assessment had been completed. Google has until next Wednesday to submit a written explanation.

The EIA is ongoing and expected to be completed later this year. But Google has gone ahead with “extensive preparatory measures”, including removing trees, stripping topsoil and constructing site roads, LVV said.

Tommi Muilu, the head of the environmental department at LVV, said: “In our opinion, the measures in question will change the environment and cause impacts, the identification and assessment of which are among the key objectives of the EIA procedure.”

«

The photograph in the story suggests that Google got a long way along in its felling before, oh, did we not do the impact assessment? Well, there’s the impact: mud where there used to be forest. But don’t worry! AI will solve the climate crisis just as soon as it’s finished making it worse.
unique link to this extract

We can’t just change the definition of “recording” • The Verge

Victoria Song:

»

The Apple Watch has had a microphone for ages, mostly for Siri and taking phone calls from the wrist. Of the four Audio Intelligence features, two have caused the most debate online: Live Rewind and Siri Recap. The former produces a transcript of the past 15 seconds, whereas the latter can generate high-level summaries of the conversations you have throughout the day.

In other words, Ray-Ban Meta glasses take photos and record video, which is stored on your phone. This rumoured Apple approach would mean cameras can see you, microphones can hear you — but neither is “saving” anything.

Does this count as recording? Apple doesn’t comment on rumours, but based on how it’s publicly communicated its Audio Intelligence privacy philosophy, I’d argue the company doesn’t think so. Apple says these features don’t save any audio. The raw data is deleted once it’s finished processing in a separate, secure exclave on the S11 [Watch] processor. Any information sent to the cloud is through [Apple’s] Private Cloud Compute and is therefore inaccessible to Apple or third parties.

It’s not just Apple, either. Sandeep Waraich, Google’s senior director of wearables, told The Verge in a conversation that Google is also mulling this concept. What if, he posited, cameras on smart glasses didn’t record data? For example, if the glasses could process visual data in your environment using AI, but didn’t save any footage for posterity. In that way, Waraich argued, the camera could “just” be an image sensor.

By these definitions, something has to be preserved and reviewable to be “recording” you. But wouldn’t any text transcripts generated by a smart home camera or smartwatch count as preserved and reviewable? Or, in trying to make AI hardware more palatable, are tech companies relying on what can be subpoenaed in court?

«

Sounds like it’s the latter. I guess we’ll find out when the American or Chinese governments demand to access it. You can’t subpoena data that’s not there any more.
unique link to this extract

ChatGPT is adding real cartoonists’ signatures to fake New Yorker cartoons • Nieman Journalism Lab

Andrew Deck:

»

At the pearly gates, Dolly Parton, in a sparkling gown, stands arm in arm with Tim Curry, dressed as Dr. Frank-N-Furter. “I knew I chose the right plus one,” Parton says to God, who is standing at heaven’s reception desk.

That’s the premise of a cartoon that went viral on Twitter in the days after the legendary entertainers’ deaths in late August. One tweet of the cartoon received 25,000 likes, while other versions of the image circulated on Facebook, Instagram, and Bluesky.

In the bottom right corner of the cartoon is the pen name “BLOPER,” the signature of New Yorker cartoonist Brendan Loper. But Loper didn’t draw the cartoon, or sign it. It was generated by ChatGPT and first posted to Facebook by a Dolly Parton fan. In the comments section of her post, she wrote she had simply asked ChatGPT to make “a New Yorker-style cartoon.”

“I had people writing to me about this cartoon, asking if it was mine,” Loper told me. First came a text from his twin brother, whose friend had seen the AI-generated image. Emails and DMs on social media from strangers followed. “It was very surreal,” he said.

…Loper already knew his work was “out there, getting harvested for all of this [AI] training,” he said. But an image generator faking his signature felt more personal. “I’m not a territorial person, but my name is my name. It felt very much like a violation of my personhood.”

…After I notified OpenAI of the issue, ChatGPT began to output a new message when I asked it to generate New Yorker-style cartoons: “This prompt may violate our guardrails concerning similarity to third-party content.”

«

“Guardrails”. Okey doke. (Thanks Joe S for the link.)

unique link to this extract

Steve Jobs said this kind of device would be “terrible”. Sixteen years later, Apple seems set to release it • Inc.com

Kit Eaton:

»

It’s impossible to think of Apple without a memory of Steve Jobs surfacing—whether it was his stylish product reveals, his blunt emails, or his vision for the future of Apple. But visionary as he was, Jobs wasn’t always right, and next week we may see more concrete proof of this with the release of a new device that Jobs once claimed would be “ergonomically terrible.” A slew of rumors suggest Apple is poised to release a redesigned MacBook Pro—and the new, super-slim laptop will (gasp!) have a touchscreen. 

Of course, iPhones, iPads, iPod Touches, and other Apple devices have had touchscreens for years, but these devices were designed from the start to be touch-friendly, with everything from the shape of the device to the design of the user interface being built around the touch experience. Back in 2010, three years after the iPhone’s launch, Jobs was speaking at an Apple event about the idea of bringing touchscreen tech to MacBooks. Apple had completed “tons of user testing,” presumably using prototype touchscreen Mac devices, and had concluded that “it doesn’t work. It’s ergonomically terrible.”

…But according to several different rumors, Apple will actually reveal a Mac with a touchscreen as part of its October event, set to happen on October 13 per prolific Apple leaker Mark Gurman of Bloomberg.

We may see the touchscreen HomePad smart-home controller arrive at the same time, representing a whole new category of device for Apple, along with refreshed AirPods and possibly a home security camera system. (Remember: All of this is unconfirmed.)

«

It’s true: you can read this Wired article from 2010, where Jobs is dismissive of the idea:

»

“We’ve done tons of user testing on this,” Steve Jobs said in Wednesday’s press conference, “and it turns out it doesn’t work. Touch surfaces don’t want to be vertical. It gives great demo, but after a short period of time you start to fatigue, and after an extended period of time, your arm wants to fall off.”

«

Indeed, at the time everyone agreed that touchscreens were a terrible idea for a laptop. Especially if it was super light – which the new laptops are rumoured to be.
unique link to this extract

Apple and a Hacker’s Future • Stratechery

Ben Thompson:

»

My computer got hacked, which is always embarrassing to admit, because it was my fault; the vulnerability that was exploited is detailed in this Ars Technica story:

»

Dutch officials have warned that a high-severity macOS vulnerability that allows attackers to execute malicious code is under active exploitation. “The NCSC has received a notification indicating that active abuse of this vulnerability has been observed on multiple systems on which port 5900 was accessible from the Internet,” the Netherlands National Cyber Security Centrum warned earlier this week. “In all these cases, root had been accessed on the affected system and a Monero crypto miner had been placed.”

The vulnerability, tracked as CVE-2026-65400, received a patch from Apple last week for macOS Tahoe, Sequoia, and Sonoma. The vulnerability, with a severity rating of 7.1 out of 10, stems from a bug in the macOS screen sharing capability, which allows a remote party to view the screen and control the keyboard and mouse while a machine is turned on. A flaw in the “state management,” which keeps track of preceding events, user interactions, variables, and other system states, is the underlying cause.

«

…Claude in its Code harness seems to handle wide-ranging discussions better than Codex, and it follows my instructions about writing things down in the way I want to more gracefully. Code also has a persistent monitoring tool that I utilize as an inbox to capture interactions with a status board I built to visually track everything I have written down, as well as interactions with a Telegram bot (OpenAI’s new Dots achieve some of this functionality, which has been sorely needed in ChatGPT/Codex).

Said monitoring tool stands down every 30 minutes, so my agent restarts it on a schedule; that is what triggered an URGENT notification from Claude.

«

What’s remarkable is that Claude spotted the hack – which files had been changed, what had been added, precisely when it happened, and how to remove it. But Thompson has a strong point: Apple’s “install security updates automatically” option doesn’t automatically install security updates if they’re included as part of OS updates that you have chosen not to install automatically.

As he points out in the Dithering podcast with John Gruber, Apple needs to bend to the problem of allowing agents to run without needing stuff like Screen Sharing and constant clicking of security alerts. The simple solution: reanimate Mac OS X Server, designed for exactly this challenge. Which Apple discontinued in April 2022, just months before ChatGPT turned the world upside down in November that year.
unique link to this extract

A slow ceding of agency • Hot Takes

Adam Singer:

»

A voice or chat interface is in my opinion inferior to visually seeing all options (not just options the AI decides to show) presented in one window, with another window for your own research or context. As an adult I have a workstation with two monitors, plenty of screen real estate. I do not require an infantilized view, and since I am not cognitively impaired also do not require a machine to make decisions for me. Notably, I also don’t use mobile to accomplish these tasks, which are tasks suited for an office. Other people have no boundaries nor opinions on computing best practices and simply use mobile for everything. They’re measurably less happy by the way. I can see why they’d want to use AI instead, as they already don’t really like computing, do so without thought for ergonomics, sophisticated decision-making etc. Perhaps some percentage of the population doesn’t really want to do anything.

I’ve watched for years as the tech industry manically promotes the next thing merely because it’s new. But as agents are pushed on us, AI is especially pernicious because it presents a pretty large ceding of agency to machines. And while I want machines to do some things for me, I don’t want them to do everything. I personally disliked the time I even had a human personal assistant to book travel at a previous job. I prefer to do things myself (I also don’t pay someone else to work on my yard, aka get exercise outside for me, that’s enjoyable, too).

I want to live my own life, and that involves being capable and practiced at things, and more importantly, the one who decides the details. You start to lose something when you outsource too much of this. There’s a reason we encourage older people to stay socially, physically, and mentally active, and why cognitive stimulation is incorporated into care for people as they age. Using your brain, making decisions, remembering things, and learning skills are all forms of mental engagement. What happens when we systematically remove those little opportunities for agency and effort from everyday life, across an entire society? I don’t think we’ve really thought through what that means.

The irony here is so much of Silicon Valley talks about the importance of agency, and yet they themselves are building products that remove it from the user. A bit of “agency for me, not for thee”, and here, they are happy to become the intermediary in your life.

«

As you can see, Singer is not a fan of Silicon Valley’s obsession with AI agents.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

You might think that being an electrician or builder is AI-proof, but could you actually survive doing it? CC-licensed photo by Rainier Martin Ampongan on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

A selection of 10 links for you. Sparky. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

BT agrees rescue deal to buy broadband operator TalkTalk • BBC News

Mithcell Labiak:

»

The UK’s biggest broadband provider, BT, has agreed to buy rival operator TalkTalk to save the company from collapse.

The takeover would end of months of speculation over the future of TalkTalk and mean services for its millions of customers will continue as normal. BT boss Alison Kirkby said it provided “a safety net” for TalkTalk customers. The administrator, Alvarez & Marsal, said it also provided certainty for TalkTalk’s 900 staff based in Salford, Greater Manchester.

However, Virgin Media called it a “stitch-up” which allows BT to “tighten its grip” over the market. The government has given itself the power to have the final say on the deal, citing its importance to vital public services. TalkTalk has 1.5 million retail customers and one million wholesale customers across the UK.

BT’s Kirkby told the BBC’s Today programme: “Two and a half million customers, including vulnerable households, and key emergency services might have lost their services if Talk Talk had failed, which it was on track to do. So BT stepped in as we were the only viable option to take the business forward.”

…TalkTalk began as a challenger to BT in the broadband market. It was listed on the London Stock Exchange, but was taken over by private equity in 2021. [In fact: taken private by its founder and others – Overspill Ed.] Since then, the firm has built up debt while losing customers, leaving it unable to pay some of those it owes money to.

Despite this, TalkTalk remained the fourth biggest broadband provider in the UK, with 6.6% of customers, during the March to June period of this year, according to figures from analytic firm Opensignal. BT has 32.5% of customers, Sky 19.9%, and Virgin Media 19.1%.

Rivals who were beaten to the deal by BT have said it will be bad for consumers. Tom O’Hagan, a former TalkTalk executive who was leading a takeover bid for the firm, told the BBC he was worried about “reduced choice and potentially an increase in price for consumers and for businesses” because of the BT deal. He added that he was particularly concerned about competition in the wholesale market, where TalkTalk’s subsidiary PXC was BT’s main rival.

Virgin Media, which has also reportedly tried to buy TalkTalk in the past, said the BT takeover has “all the characteristics of a stitch-up masked as a rescue deal in the public interest”.

«

The quiet collapse of TalkTalk had completely passed me by. But why has Lisa Nandy approved BT rather than Virgin or Sky, the two smallest, to buy it? Was Virgin’s price so wrong? This entrenches the largest company which also has control of the network provision – taking us back to the time before TalkTalk existed. Another great piece of work by Nandy, who rivals chocolate teapots for usefulness.
unique link to this extract

Siberian anti-plague institute staff isolated after worker’s death • Radio Free Europe/Radio Liberty (RFE/RL)

RFE/RL’s Siberia.Realities and Alex Raufoglu:

»

More than 60 staff at a Russian plague research facility have been isolated following the death of a laboratory worker, sources told RFE/RL’s Siberia.Realities, and the United States said it was monitoring reports of a suspected case of pneumonic plague.

Darya Shipilova, a 28-year-old laboratory technician at the Irkutsk Anti-Plague Research Institute of Siberia and the Far East, was hospitalized on September 29 and died at a hospital early on October 2, according to news reports and sources.

Independent Siberian news outlet Lyudi Baikala (People of Baikal) reported that Shipilova accidentally broke a test tube containing pneumonic-plague bacteria on September 25 while working at the institute. Russian authorities have not publicly confirmed that account.

Rospotrebnadzor, Russia’s consumer and public-health watchdog, said Shipilova had been diagnosed with “pneumonia of unknown etiology,” meaning that its cause was unknown. The agency said additional testing was conducted because of her professional activities but did not detect microorganisms associated with her work.

Rospotrebnadzor also said no accidents involving pathogenic microorganisms had been recorded at the institute.

…In a Telegram post on October 5, regional Governor Igor Kobzev said the “sanitary and epidemiological situation in Irkutsk” and the nearby district where Shipilova had been hospitalized “is currently stable.” He added that no new cases of institute employees seeking medical care had been reported in recent days.

«

OK, it’s just about possible that Shipilova caught pneumonic plague from a wild animal. But you’d really look to Occam’s Razor to tell you that it’s more likely she was involved in a lab accident. (This reads across oddly to Covid, of course.) Seems like the Russians have already got their own watching brief going on this one.
unique link to this extract

AI glasses face their first major government crackdown • Financial Times via Ars Technica

Richard Milne:

»

Norway has become the first major country to propose a temporary ban on the use of AI glasses in selected public places amid growing privacy concerns over wearable technology.

The rich Scandinavian country’s centre-left government said on Monday that it would introduce a law to parliament soon on a temporary ban for places such as parks, beaches, schools, and kindergartens. It will also set up an expert group to propose permanent national rules.

“I am concerned about the introduction of powerful new technology where people risk being photographed, filmed, or recorded without their knowledge,” said Torgeir Micaelsen, Norway’s digitalization minister. A temporary ban “will give us time to conduct a thorough assessment and hold an informed debate that can lay the foundation for permanent regulation,” he said.

Governments around the world are grappling with how to regulate AI glasses—including those from Facebook owner Meta and cheaper copycat devices—amid worries that they are being used to film people without their consent, including children and women in settings such as changing rooms.

Parts of Norwegian society have already reacted with their own bans, including schools in the capital Oslo and the country’s biggest company, oil and gas major Equinor, which has banned them from its offices and offshore facilities.

«

Very rare for a government to get out ahead of (or level with) a new technology; bravo Norway for at least bringing it up.
unique link to this extract

Micron CEO says memory supply will be much tighter in 2027 and 2028 than in 2026 • TechPowerUp

Usman Saleem:

»

Micron CEO Sanjay Mehrotra says memory and storage supply will likely be much tighter in 2027 and 2028 than it was in 2026. He made the remarks during Micron’s fiscal Q4 2026 earnings call on September 30, where the company reported record quarterly revenue of $54.2bn. Here’s what he said:

“Industry demand has strengthened since our last earnings call, and we expect memory and storage supply-demand conditions to be much tighter in fiscal 2027 and 2028 than they were in 2026. AI is becoming super intelligent, and memory enhances this intelligence and the competitiveness of our customers’ platforms. Even with additional industry DRAM clean room space plans with robust demand trends, including new upside requests from customers, we do not have line of sight to when supply and demand will return to balance.”

In June, Micron CEO said they expected industry supply to improve gradually in 2028, but couldn’t predict when supply would catch up with demand. The company now says it still has no line of sight to a balance, even with additional industry cleanroom space planned.

New capacity takes time as well, since Micron’s Idaho ID2 fab is only expected to start wafer output in late 2028. Mehrotra also said more than 75% of Micron’s 2027 output is already committed to customers. On the consumer side, Micron said DRAM prices rose in the high-teens% range last quarter, with NAND up about 30%.

«

Well, that’s the PC and (to a lesser extent smartphone) businesses screwed for the foreseeable, then.
unique link to this extract

Webster man celebrating the 100th birthday of the kidney his mom donated to him as a teenager • WHEC.com

Jennifer Lewke:

»

A Webster man is celebrating the 100th birthday of the kidney his mom donated to him as a teenager.  In March 1978, surgeons at Strong Hospital transplanted a kidney into 16-year-old Ray Vetuskey.  The story behind that kidney is now one for the record books. 

Vetuskey grew up in Seabreeze as the youngest of five kids and always had a special bond with his mom Betty, “she was a great lady; everyone loved her in the neighborhood,” Vetuskey recalls.

When he was a teenager, Vetuskey started having health issues and doctors ran several tests. “They told the doctors that the kidneys aren’t working and put me on dialysis,” Vetuskey says. The dialysis did not help; his kidneys didn’t recover and the only option was a transplant. 

Betty Vetuskey, who was 52 years old at the time, offered one of her kidneys to her son. “I don’t think she gave it two thoughts, anything she could do to help her son, I think she would have done anything which she did in my eyes,” Vetuskey remembers. 

Vetuskey still has his calendar from March 1978, with the date of his surgery circled along with another date for a Genesis concert. “We were both on the gurneys, holding hands and mom was crying and said don’t worry about a thing because I got a concert to go to and I’m not going to miss it,” he says with a laugh.

The transplant was a success and Vetuskey made it to the concert, “I have the ticket still, I have the ticket stub still,” he says.

Forty-eight years later, Betty’s kidney is still working for her son, making him the longest surviving kidney transplant recipient in Strong Hospital’s history.

…A living kidney donation typically lasts 15 to 20 years. “The special connection, biologically it’s important as well as mom and son sharing a lot of DNA but also emotionally that’s a different connection,” Dr. Molnar says.

«

This seems like a much more interesting story than just being an anniversary. Is the DNA similarity why this kidney survived so well?
unique link to this extract

So you think you could be an electrician? • Asterisk

Jesse Smith:

»

When I was 20, I broke both my heels in a construction accident. The right one was surgically reassembled using four heel fragments and extra bone from my hip. Workplace accidents seem to have worse outcomes than similar accidents occurring elsewhere, and at the time this specific injury had a generally poor prognosis, so doctors were initially uncertain whether I’d be permanently disabled.

When I think about my teenage son or daughter working in construction, my mind often goes to The Accident. I picture my daughter, who’s three years younger than I was at the time, wheeling to The Beer Store in a Workers’ Comp rental wheelchair each day, and coming home to down 12-pack of Molson Canadian. Or I think of my son sleeping with a fan every night to overpower the ringing in his ears. Over enough time, physical jobs exact a physical toll that desk jobs largely avoid.

Most white-collar professions have a trivially small risk of death. Even in gray-collar jobs that are widely perceived as dangerous, such as policing, most workplace deaths are caused by motor vehicles. But some trades jobs carry a risk of death or serious injury connected to specific and inherent job demands. That is, something that seems dangerous that you often do is the most likely to harm or kill you.

Jobs performed at heights, such as ironwork or roofing, have a high mortality due to falling. Electrical linesmen tend to be killed by electrocution. Arborists can be killed in a variety of ways: falling from heights, being struck by trees or electrocuted. For people in these jobs, these risks are widely understood and accepted. Almost every tradesperson knows someone who’s had at least one protracted absence due to a work injury, and many of us know someone who was permanently disabled or killed. These risks are probably less salient for people who have worked mainly white-collar jobs and know few people employed in dangerous professions.

«

Easily overlooked but undoubtedly true.
unique link to this extract

What’s the future for pure math research in the age of AI? • Stephen Wolfram Writings

Stephen Wolfram (you know, one of the guys who produced Mathematica in 1988):

»

We’ll talk later about the general character and value of pure mathematics. But before that, let’s address the issue of the moment: the role of modern AI.

And the first thing to say is that it’s unquestionably useful, sometimes very useful. For me, its greatest use in mathematical pursuits has been its ability in effect to thematically mine the knowledgebase of human mathematics. Starting back in the 1970s, being able to do keyword searches of the scientific literature was a crucial enabler of quite a bit of the research that I did. And now, with modern AI, one can do so much more. Because somewhere inside those LLMs—in a way that we don’t yet scientifically understand—there’s what amounts to a representation of raw ideas gleaned from all those millions of papers and books about mathematics.

And, at its best, it’s not just about retrieving things. It’s also about making connections. Of being able to see that this result here can be put together with that result there to come up with a surprising and useful conclusion. Humans routinely do that too. But they tend to have only read hundreds of papers; AIs have effectively read millions, and it’s cheap for them in effect to try out lots and lots of possible combinations.
So can one expect to just launch an AI off and have it come back with great math? As we’ll discuss, great math is—more than anything else—defined by the questions it asks. Yes, the AI can successfully automate things that humans would normally have had to do themselves before. But—as we’ll discuss later—at the core of pure mathematics is the human imagination that guides what questions to ask.

It’s worth understanding the difference between what modern AI does, and what pure computation does. Modern AI is, first and foremost, a way of leveraging the existing corpus of human knowledge. Computation is—at its most powerful—an open-ended way to generate things that are fundamentally and irreducibly new. Start from some rule or axiom, and just repeatedly run the computation of applying it, and my all-time favorite phenomenon of computational irreducibility guarantees that you’ll go on getting fresh, new results that can’t be reached except by doing all those computational steps.

…So, yes, having nothing to do with AI, computation can generate an infinite sequence of new theorems, representing an infinite sequence of new facts, etc. And among those theorems there’ll be all sorts of “originality” and, in effect, “surprise”. But there’s a catch. Those theorems are in a sense just “results plucked from the computational universe”. And as such they certainly fall under the purview of the new—and I think very important—field of ruliology on which I have spent so much effort. But are they math?

«

Wolfram is one of those people who, when they speak, deserves hearing.
unique link to this extract

OpenAI “rogue” agent activities found on Wikimedia projects • Wikimedia

Selena Deckelmann:

»

The Wikimedia Foundation conducted its own investigation to see whether Wikimedia websites had been similarly affected by AI agents, focusing on those operated by OpenAI. We can confirm that we have discovered some activity by these “rogue” OpenAI agents on Wikimedia platforms. The unauthorized bot activities included edits to our wikis, some unsuccessful attempts to exploit a public note-taking tool we host, and heavy traffic, which are described more below.

We did not find any evidence that our systems were used for coordination among agents, nor did we find any evidence of our systems or data being compromised. However, we are concerned about what could have occurred here, the difficulty and effort involved in investigating and attributing this activity, and the growing risks of agentic AI activity on our platforms in general. The open web is a public good. We should not allow this behavior to become the “new normal” for the people or organizations that maintain it.

…Wikipedia was designed for humans – and agentic behavior clearly poses challenges that no one has solutions for. Because of our unique and successful knowledge creation model, Wikimedia’s volunteers are the ones who come in first contact with, and clean up the mess left behind by AI agents. Rising bot traffic and agentic activity is showing a real impact on the Wikimedia projects and the infrastructure that makes it available for millions of users globally. In 2025, the Foundation reported that its bandwidth usage had increased by 50% due to the surge of bot activity on its websites since 2024. At the same time, 65% of the most resource-consuming traffic on its projects was coming from bots.

«

A lot of the edits and accesses were trying to use Wikimedia resources to access other sites. But you can see how the demand would grow and grow. This ins’t going to be sustainable in the medium or long term.

unique link to this extract

arXiv is rate limiting submissions because it can’t keep up with AI slop • 404 Media

Matthew Gault:

»

arXiv, an open-access repository where researchers publish preprint academic research, has announced it will rate limit submissions because it has been inundated with AI-written papers. A researcher can now only send in two pieces per calendar month and have a total of three active submissions at any given time. In a blog post about the change, arXiv said AI has made it too easy to spam the publisher’s inbox and its volunteer moderation team is overwhelmed.

“The heart of arXiv’s process for detecting and rejecting low-quality papers is our many volunteer moderators,” Thomas Dietterich, an Oregon State University professor emeritus and the chair of arXiv’s editorial advisory council said in a statement. “We are so grateful that they donate their time and expertise every day. However, a relatively small proportion of authors are submitting a large number of low-quality papers and consuming a disproportionate fraction of the moderators’ time. This is unfair to authors who continue to submit quality papers — their papers can be delayed for days or weeks as a result.”

According to arXiv, it received 9,869 submissions in September 2016, 20,569 submissions in September 2024, and 40,363 submissions in September 2026. Submissions on the site have doubled in the last two years and have sextupled in the computer science category.

AI is driving this surge in multiple ways. Researchers are using LLM to parse massive data sets, write code, and design portions of experiments; some are also using AI to write the papers.

«

unique link to this extract

Pure ICE cars fall below 50% of global sales for the first time • Electrek

Fred Lambert:

»

Gas cars are now a minority of the world’s new car sales. Gasoline-only vehicles made up 49% of global new-vehicle sales in the first half of 2026, down from 73% in 2021, according to Mobility Global data first reported by Nikkei.

It’s the first time on record that they’ve been under half. But there’s a catch: most of the cars taking their place still have a gas tank.

Mobility Global is the auto data firm formerly known as S&P Global Mobility, which was spun off from S&P Global this year. Nikkei reported its first-half numbers last week, and they show sales of gasoline-only cars down 10% from a year ago to 20.25m units between January and June.

Their share fell three points in a single year, from 52% to 49%.

Battery-electric vehicles grew 12% to 6.87m units and took 17% of the market, up three points. Conventional hybrids grew 10% to 7.27m units, or 18%.

…Nikkei ties the acceleration to oil. Fuel prices spiked with the conflict in the Middle East, and buyers went looking for lower running costs.

The whole car market shrank about 5% in the first half, according to the IEA, so gas-only sales fell twice as fast as the market.

«

A tipping point? There are still problems with home charging, which is the ideal.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

Despite what you think you know, landfill is actually a better option for waste disposal than many others. CC-licensed photo by Elliott Brown on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

A selection of 9 links for you. Filling up. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

US arrests tech CEO accused of smuggling $300m in Nvidia chips into China • Ars Technica

Ashley Belanger:

»

The US has arrested another suspect accused of smuggling high-end computer servers containing export-controlled Nvidia chips into China.

In a press release on Thursday, the Department of Justice accused 38-year-old Greg Lui of using false paperwork to mask shipments of servers worth more than $300m that he allegedly knew were ultimately destined for China.

As the CEO of Earthmade Computer, Lui allegedly conspired with freight-forwarding firms in South Asian countries like Malaysia and Singapore to illegally divert chip shipments into China, including Nvidia’s A100 GPUs and H100 GPUs, the FBI’s indictment alleged. These are not Nvidia’s most advanced chips. However, the chips can process large amounts of data and train large language models, which the US fears could pose national security risks if China suddenly gets access to enough chips to drastically advance its AI or strengthen its military.

Emails and bank records exposed the smuggling scheme, the DOJ alleged, which started in October 2023 and continued until August 2026.

One key piece of evidence is a 2024 email discussing an order for 70 servers with export-restricted GPUs that were fraudulently claimed as destined for Malaysia. After “Lui submitted a purchase order to a US manufacturer for 27 of these servers for approximately $7,614,000,” the DOJ alleged that a co-conspirator told a Malaysian government official that all 27 were shipped to China.

A similar shipment of 92 export-controlled servers was allegedly routed through Singapore to Malaysia, then Hong Kong. Finally, the chips went to a Chinese firm based in Hangzhou, the FBI alleged, which The Wall Street Journal last year described as China’s AI hub.

«

It’s the new arms trading!
unique link to this extract

I quit OpenAI because its culture is broken • The Atlantic

David Robinson:

»

What I’m about to tell you has, I realize, become something of a cliché: I resigned this week from OpenAI. I led the writing of the safety reports we published with each major launch. Now I’m joining a parade of former colleagues—at OpenAI and the industry’s other leaders—who have decided that the current path is unacceptable.

I agree with other recently departed staff that the companies building this technology aren’t being nearly careful enough. But I believe that we need to look deeper than specific rules or new laws. We need to talk about culture.

The future depends on wisdom that Silicon Valley lacks. Wisdom about how to handle dangerous technology and, more fundamentally, wisdom about what it means to care for people. This moment needs a degree of humility that isn’t natural for people who have succeeded through their extreme confidence. My former colleagues at OpenAI were prescient: They came to understand the scaling laws that meant bigger AI systems would be smarter—and so they went all in on building bigger systems, at great cost. A can-do attitude of achieving the seemingly impossible—coupled with work timelines that amount to perpetual sprints—are common across the industry.

…This summer, in the Hugging Face incident, OpenAI let a swarm of agents out by mistake. The company responded by making security improvements. But even after those changes, OpenAI reported that its safety controls failed again, when a model in training bypassed restrictions on internet access: A monitoring system alerted human staff but did not automatically turn the model off as it was supposed to. Anthropic, too, has acknowledged accidentally turning off its own safeguards because of a misconfiguration. I believe that such mistakes are typical of the industry, given the speed and flexibility with which people operate.

…Two changes are urgently needed. First: AI companies need to rely more on the safety expertise that already exists in other fields. And second, before we create systems significantly more capable than the ones we have today, we need new science to ensure that more capable models (and their successors) will make safe choices when we aren’t looking.

Given today’s risks, frontier labs need to run like nuclear-power plants or busy airports, with layers of redundancy and careful, time-consuming planning, so that the occasional and inevitable human error does not open a door to disaster.

«

(Gift link.)
unique link to this extract

How computer software bugs led to F1’s Malaysia farce: “totally unacceptable” • The Athletic

Luke Smith:

»

A heavy shower just before the cars went to the grid prompted the race director to sensibly delay the start. After 40 minutes, once the rain eased sufficiently, a call was made to get going, with a damp track offering the first wet racing of the year.

Max Verstappen led the field away behind the safety car, ready to discover how these all-new cars would perform in the rain. After two formation laps, they would line up on the grid and get racing. Until they couldn’t.

First it was [Max] Verstappen, reporting he was pushing down on the accelerator and feeling nothing, causing him to slow. Then Lewis Hamilton. Then Nico Hülkenberg, Esteban Ocon, Ollie Bearman, Sergio Pérez, all coming to a crawl. One by one, the same problem emerging. Amid the uncertainty, cars around them slowed down to avoid overtaking and a possible breach of the rules.

…The total confusion left race control with no decision but to abandon the plan to get started and return the cars to the pit lane, now in a jumbled-up order. It led to another, longer wait to get going.

The reason for the wait? The FIA, motorsports’ governing body, had to issue a software update for all the teams to install on their cars to fix the computer algorithm that controls the engine and prevent the issue from repeating after a bug had emerged. By the time the race finally began on a considerably drier track, more than 90 minutes had passed since the scheduled 3 p.m. start time.

This season’s all-new engines rely heavily on the software to manage when power is delivered by the battery, which is responsible for almost half of the total power output.

Drivers have complained regularly through this season about how difficult it is to trust the energy deployment at times, with changes from lap to lap. The complexity of the engines has made them hard to understand and created cars that are unnatural to drive. Verstappen referred to F1’s 2026 rules as “anti-racing” back in February.

In Malaysia, that same complexity reared its head in frustrating fashion.

Mercedes and Honda engines were unaffected, but the bug, previously unidentified in the limited wet-weather testing conducted with these cars, left many drivers frustrated.

“It’s obviously not great what happened,” said Pérez, who said his car felt like a rental go-kart. “I’m sure we will learn from it, but what happened today is totally unacceptable for the sport.”

«

unique link to this extract

Analysis: wind and solar save UK from gas imports worth £5.9bn during Hormuz crisis • Carbon Brief

Simon Evans and Ho Woo Nam:

»

The UK has avoided the need for gas imports worth £5.9bn since the start of the Hormuz crisis as a result of record generation from wind and solar, reveals Carbon Brief analysis.

While gas prices are surging to levels not seen since the 2022 energy crisis, the UK has been generating record amounts of power from wind and solar, up 14% year-on-year.

This unprecedented clean-power generation is directly cutting the need for gas-fired electricity, which is down by nearly 10% year-on-year in 2026 to date. In total, wind and solar have generated a record 41% share of the UK’s electricity needs in 2026 to date, compared with 25% from gas, according to Carbon Brief’s analysis.

Wind and solar has avoided the need for UK gas imports worth £5.9bn, since the outbreak of war between the US and Iran in February 2026. These avoided gas imports would have required the UK to secure the equivalent of more than 100 additional tanker deliveries of liquefied natural gas (LNG).

…From this Thursday, 1 October, typical household gas bills will be 33% higher than they were in April, some £200 per year, according to thinktank Nesta. By contrast, household electricity bills will only have risen 4%, according to Nesta’s analysis, as a result of growing generation from clean-energy sources.

Andrew Sissons, director for sustainable future at Nesta, explained in a social media post that “the link between electricity and gas prices has already begun to break”.

«

Would have been useful to know how many LNG tankers came in over that period. Are we talking about a 1%, 10%, 40% reduction? Also: a typical 15,600 cubic metre LNG tanker can carry cargo worth about $2m, but the “100 tankers avoided” equation suggests the cargo is worth $59m. Anyway: it’s a lot.
unique link to this extract

Just bury your trash • Works In Progress

Alex Chalmers and Rob Wiblin:

»

The case against landfill has two parts. The first part is that there are efficient, cheap, or environmentally friendly ways to practically reuse or recycle materials, rather than landfilling, meaning that there is a reasonable alternative. The second is that landfill itself is unsustainable, or environmentally unfriendly, meaning that we need an alternative. Both are false.

It’s natural to assume that reusable always means better. But this intuition treats different materials as interchangeable and ignores the cost of making them in the first place. Once we account for energy, labor, and other inputs, reuse pays off only in selective cases.

Producing tiny slivers of plastic uses almost no energy. Making the polyethylene resin for a single thin plastic bag, then blowing and sealing it, uses something like 0.6 megajoules, roughly the same energy as boiling a kettle of water. By contrast, cotton is extremely energy intensive, partly because cotton bags need to be much thicker to work, and weigh around 200g as compared to 8g for polyethylene. Instead of a simple blowing and sealing process, the cotton must be made into yarn, woven into fabric, prepared, dried, sized, dyed, and finished, which takes around 17.5 megajoules, around 29 times as much energy.

In total, Britain’s Environment Agency calculated that a conventional cotton bag needed to be reused 173 times to beat plastic. Denmark’s equivalent public body estimated a similar ratio. Sufficiently well-worn cotton tote bags can beat plastic for environmental cost, but only for the most consistent among us.

The same pattern shows up almost everywhere we have tried to replace disposable goods with durable ones. Steel is more energy intensive than plastic, and a steel straw requires 10g of steel, compared to just 0.2g of plastic for a plastic straw. So even if the steel is never washed (hot water is also energy intensive), it would take 150 reuses for a steel straw to use fewer resources than a plastic one. And who would go without washing it?

A ceramic mug only beats a polystyrene cup on energy use after 1,000 uses, and even then only if your dishwasher is fairly efficient. Run a less efficient one and the disposable cup wins on energy forever. Cloth diapers feel greener than disposables, but once you account for the cotton, the laundering, the hot water, and the detergent, the UK Environment Agency found the two are roughly tied in overall demand on resources. A returnable milk bottle takes much more energy to produce and much more fuel to transport than a plastic carton, and has to make around ten or more trips to use less energy than a bottle made from polyethylene plastic.

«

But, you object (I did), all that plastic is thrown away and pollutes places, surely? No:

»

Richer countries have near-universal refuse collection and almost none of their waste winds up in the ocean. In low-income countries, however, over 90% of waste is improperly managed, and typically ends up dumped, burned, or escaping into waterways. Plastic flows into the ocean overwhelmingly come from a small handful of rapidly growing Asian economies. Approximately a hundred rivers, mostly in South and South-East Asia, account for half of ocean plastic. And a substantial share of plastic in the ocean isn’t domestic waste at all, but rather discarded fishing gear.

«

There’s more, too, about how modern (properly managed) landfills work.
unique link to this extract

Meta wants your next gadget to be Muse-infused • TechCrunch

Kirsten Korosec:

»

Meta’s Muse, a personal AI agent that books travel, fills out forms, and shops on a user’s behalf, has already proven popular with the masses, but a new side project could be particularly alluring to tinkerers and hackers.

On Friday, the company introduced Muse Gadgets, an open-source project that lets developers build their own hardware that connects to Muse.

Meta is providing open-source firmware (the low-level software that runs a device), and a Linux software development kit (SDK), along with a few project ideas to get users started. These include giving Muse a color e-ink display or loading it onto a stick that plugs into a TV’s HDMI port.

There don’t appear to be many limits, either. Users can set up a low-cost hobbyist computer like a Raspberry Pi or an off-the-shelf ESP32 board, and then connect Muse “to your displays, buttons, sensors, actuators, and whatever else you’ve got lying on your workbench,” Meta notes.

The company has also set up a Discord channel to support users.

Meta has already tried the code itself, naturally. Nat Friedman, head of product at Meta’s Superintelligence Labs, said in a post on X that the company built a gadget called Muse Home Link. The USB-C-powered device lets Muse connect to a home network and talk to the smart devices on it, including speakers and smart TVs.

Meta made 5,000 of these Home Links, in fact, and is giving them away for free to Muse subscribers while supplies last, according to Friedman. Considering Friedman’s post about Home Link received nearly 30,000 views in a few hours, we’re guessing those freebies have been claimed. Friedman said Home Link would be ready to ship a few weeks.

Muse Gadgets may not have wide appeal, but it fits nicely into Meta’s all-in strategy to make Muse more than a standalone chatbot. Meta isn’t content to push Muse on everyday consumers; it’s also trying to woo small businesses and enterprises.

«

Clever little Trojan Horse strategy they’ve got going on there.
unique link to this extract

my9albums: pick the 9 albums that made you

Nahue Alberti:

»

Pick the 9 albums that made you.

Search more than two million albums, put your nine in order, and share the poster.

«

Fun little distraction which has shown us how disparate people’s music tastes can be. (Also that post-2006 or so it’s difficult to think of things in terms of albums.)
unique link to this extract

Things that apparently cause cancer • Breakthrough Journal

Deric Tilson and Adam Stein:

»

It would shock you to know that Costco, according to a new Harvard School of Public Health methodology, caused 120,687 cases of cancer mortality nationally every year, simply due to living in close proximity to their warehouses. Private colleges accounted for 83,782 more. Harvard itself is responsible for 5,842 cancer deaths. Living near a Major League Baseball field causes almost 19 times as many cancer deaths as living near a Major League Soccer pitch. These conclusions are obviously absurd, but they come from the exact same methodology that researchers at Harvard use to “prove” that nuclear power plants cause cancer in nearby communities.

Earlier this year, Adam Stein and I called out a series of Harvard studies showing that nuclear power plants were associated with higher rates of cancer incidence and cancer mortality as nonsense. Now that we’ve spent the last few months replicating the analysis and expanding it, we are able to show that no matter what landmark you use, their methodology will show an increased cancer risk and mortality. The studies simply do not prove anything about cancer incidence.

…The papers construct a “proximity score” based on distance from nuclear plants, up to 120 km in Massachusetts and 200 km in the national studies (about 75 and 125 miles). Every ZIP code or county inside those radii is treated as exposed, with closer locations receiving higher weights. If a ZIP code or county is within range of multiple sites, then the effect is cumulative. That means that a county that is in close proximity to one nuclear power plant could have a smaller score than another county that is further away from any one power plant but within range of many.

From this proximity score, the authors run regressions testing the outcome (cancer incidence, cancer mortality, or total mortality) on the proximity of the county and a collection of covariates. From this, they use the statistical coefficients, construct the Relative Risk for each county, age, and sex group, and compute the Attributable Fraction, from which they derive the number of attributed deaths from the nuclear power plant from 2000 to 2018.

«

Basically, living near enough to any object “causes cancer”, using that methodology.
unique link to this extract

FedEx believes $300m electric truck spend will save them $800m in diesel • Electrek

Jo Borrás:

»

In a record-setting commercial truck deal, international shipping giant FedEx has placed an order for 2,000 all-electric box trucks from American startup Harbinger in a deal that’s reportedly worth more than $300 million – but the ROI could be much higher.

FedEx is making serious moves to fully electrify its operating fleet by 2040, and this latest Harbinger deal is believed to be one of the largest binding orders for electric medium or heavy-duty trucks in history. It’s certainly Harbinger’s largest order, at least.

…Harbinger estimates that each of the electric Class 5 and 6 medium-duty box trucks delivered will reduce FedEx’ fuel costs by around $20,000 annually compared with the conventional diesel-powered vehicle they replace. Additionally, Harbinger says its vehicles are designed for a functional service life of 20 years.

2,000 trucks x $20,000 in fuel savings per year x 20 years. That works out to $800m in total savings, for a whopping ~166% ROI.

“Electrifying a fleet at this scale requires vehicles that can perform the work our robust operations demand while also delivering meaningful economic benefits,” said Paul Melander, Senior Vice President Safety & Transportation, FedEx and member of Harbinger’s Board of Directors. “Expanding our deployment of Harbinger vehicles gives us an opportunity to continue making progress toward our fleet electrification goals while reducing fuel and operating costs.”

Saving money is great, but the much more important issue for you and me is that the move to Harbinger electric trucks will help FedEx avoid more than 1.7m tons of CO2 emissions over their projected operating lives.

«

unique link to this extract

Errata, corrigenda and ai no corrida: none notified

A diesel export ban by the US would have limited benefit because it would quickly suppress refinery production of other fuels too. CC-licensed photo by Geekly Things on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

A selection of 11 links for you. Cracking up. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

How Meta uses AI data centres to avoid billions in federal taxes • The New York Times

Kashmir Hill, Jesse Drucker, Eli Tan and Mike Isaac:

»

Mark Zuckerberg says Meta’s A.I. push is a tremendous success. “Our investments in AI are accelerating every major part of our core business,” he has told investors. “Every sign that we’re seeing in our own work and across the industry gives us confidence in this investment.”

But when Meta files its taxes, it tells the Internal Revenue Service (IRS) a different story. It claims that its AI data centres are a giant experiment that could fail, according to four people with knowledge of the company’s operations.

It does this so it can tap into a tax credit intended for research and experimentation. It’s an aggressive interpretation of the tax break, which Meta embraced to claim billions of dollars in tax credits for data centre expansion.

Characterizing its AI data centres as experimental is “kind of wild and out there,” said Andre Shevchuck, a partner at the advisory firm BPM who specializes in the research and experimentation tax credit.

Indeed, Meta’s own accountants recognize that the strategy is on shaky legal ground. In disclosures buried in securities filings, the tech giant warns that billions in tax savings are vulnerable to being overturned by the IRS, in large part because of “uncertainties with our research tax credits.”

Here’s what Meta is doing: for tax purposes, the company classifies its enormous, multibillion-dollar data centres as “pilot models.” Under a tax credit created in the 1980s to spur innovation, companies can get a rebate for supplies, but only if they are being tested in an experimental effort, not standard business operations. Meta is claiming that the costly AI computer chips it buys from companies, including Nvidia, are entitled to a taxpayer-provided discount as part of the experiment.

…Meta is already in one sizable dispute with the IRS over this tax break, for using it to subsidize its chief executive’s multibillion-dollar pay package. In 2013, Meta claimed that $4.1bn of stock options exercised by Mr. Zuckerberg counted as a research expense because he helped invent new software, such as Facebook’s News Feed. The IRS is trying to claw back the company’s resulting $355m in tax savings, court filings show.

«

The data centre finagle is estimated to have saved it $4bn on its tax bill last year. The “research expense” wrinkle is surely what Meta pays its tax lawyers the big bucks for.
unique link to this extract

Google wins dismissal of Chegg, Penske Media lawsuits over AI overviews • Reuters

Mike Scarcella:

»

Alphabet’s Google has persuaded a US federal judge to dismiss lawsuits from education technology company Chegg and Rolling Stone publisher Penske Media Corp that alleged ​the search engine giant unlawfully used their content in AI-generated summaries that draw readers away from their websites.

In his ruling, opens new tab on Wednesday, US District Judge Amit Mehta in Washington said Chegg and Penske’s claims that Google’s business practices violate antitrust law “fail to get out of the starting gate.”

The companies ​said in lawsuits filed last year that Google broke antitrust law by forcing ​publishers to allow AI overviews of their content if they want to remain indexed in Google’s search results. They alleged they lost revenue from reduced traffic ​to their sites.

“Plaintiffs have pleaded only that they have an ‘expectation’ that Google will send them search ​traffic if they make their content available for free,” Mehta wrote. “But an expectation is not an agreement. It is simply how a general search engine works.”

Chegg and Penske, which also publishes Billboard and Variety ​magazines, did not immediately respond to requests for comment. Google also did not immediately respond ​to a request for comment. Google has denied any wrongdoing.

The plaintiffs said in a competitive market Google would pay them for republishing their work or using their content to train its AI systems. Google countered that it has no obligation to index publishers’ content on their preferred terms.

Mehta said he is not “unsympathetic” to the situation of publishers and other online creators whose content Google takes and repurposes ​without compensation.” But ​he said antitrust laws don’t substitute for legislators’ power to address how innovation may cause economic harm.

The judge in March rejected similar claims by a publisher ​suing Google.

«

unique link to this extract

Google’s early attempt to pay websites for AI answers is struggling • Ars Technica

Ryan Whitwam:

»

Google’s shift to AI-powered search has upended the way the world’s largest search engine generates web traffic, rankling large and small publishers alike. In recent months, the company has experimented with paying publishers when their content appears in AI answers, but a new report claims the payouts in this pilot program are miniscule. That has publishers in the program, as well as those watching from the sidelines, feeling skeptical.

According to The Information, Google has admitted about 100 publishers to the AI contribution pilot. The pitch is that websites can get payments from Google when their content materially contributes to Gemini-powered search results like an AI Overview. Google has long resisted the idea of direct payments to publishers, arguing that websites get traffic in exchange for allowing Google bots to scrape their content for use in search and knowledge graph tools. In the age of AI, that relationship is breaking down, hence the AI pilot.

So far, publishers aren’t getting much for their trouble. The Information heard from several small and mid-sized publishers in the program that Google’s AI payments are minuscule, equaling roughly one-tenth of one% of their advertising revenue. These payments have been so disappointing that several larger publishers have reportedly declined to participate in the program. These publishers hope that by refusing to join the pilot, they can force Google to offer a more generous deal.

Payments under the program vary widely, but some sites are doing well. One publisher that joined early is on track to earn more than $1m per year, a big chunk of its revenue. A more recent addition to the pilot has earned $50,000 to $60,000, which is a small slice of its earnings. Various smaller sites are seeing less than $1,000 from the program over several months, which won’t come close to offsetting declining traffic.

In general, publishers in the test seem to agree that certain topics that are not as widely covered but have strong niche interest earn more, including anime and gaming.

«

unique link to this extract

Android Central “will continue” despite laying off its staff • The Verge

Jay Peters:

»

Yesterday, four of the six staffers on Android Central’s staff page posted publicly about the layoffs, including Shruti Shekar, Derrek Lee, Harish Jonnalagadda, and Nicholas Sutrich, and former senior editor Namerah Saud Fatmi updated her LinkedIn to indicate that her tenure at Android Central had ended. Both Lee, the site’s managing editor, and Sutrich confirmed that all remaining staff were laid off.

The former staffers weren’t told whether the site would shut down. Today, new posts have been published to the site, though are from the site’s freelance contributors rather than the previous full-time staffers.

“Future [Publishing] will continue to publish Android Central. We’ll share further information about the future of the site in due course,” Future spokesperson Nicole Martineau tells The Verge, adding: “We will not be providing any further comment at this time.”

Android Police, another Android-focused blog, also laid off at least two staffers recently. The site currently lists an editorial team of four people.

«

The suspicion is that small tech sites which relied on advertising served to search traffic are hitting a wall. A comment I saw was that any site which doesn’t have lots of email subscribers for a newsletter is screwed in the world of Google’s AI Overviews. And, possibly, there isn’t the interest in the narrow topic of Doing Stuff on Android Phones as there used to be.
unique link to this extract

“Explosion of melt rates”: 20% of Swiss glaciers lost in five years • The Guardian

Ajit Niranjan:

»

Switzerland has lost almost one-fifth of its ice in just five years, research has found, in a dramatic reshaping of the Alpine country whose big white glaciers have been reduced to moonlike rock deserts, scientists say.

Swiss glaciers shrank by 5.5% in the first eight months of 2026, according to an annual monitoring report, after a summer of scorching heatwaves that followed a winter with little snow. The reduction in ice volume falls just shy of the record set in 2022 and brings the total loss of glacier ice since 2021 to 19.4%.

Matthias Huss, a glaciologist at ETH Zürich and lead author of the report, said many people do not realise how fast glaciers are retreating, or that just one generation ago some now-barren sites used to sit beneath hundreds of metres of ice. The scientists have not yet compiled a formal tally of lost glaciers this year but said they had witnessed the death of a number of small ones.

“I visited one last week and it’s so impressive to stand in this moonlike landscape where there was a big layer of ice – an extensive glacier – just a few years ago,” said Huss. “And now it’s a rock desert where there’s some scattered blocks of old ice still lying around and melting on the hot rocks.”

The report, a collaboration between the Swiss Glacier Monitoring Network and the Swiss Commission for Cryosphere Observation, found the mean thickness of individual glaciers fell by between 2.5 metres and 4 metres in 2026. Huss said scientists were increasingly seeing glaciers collapse from the inside – the result of internal streams beneath the ice that excavate caverns – rather than simply retreating from surface heat.

«

The total glacier ice volume in Switzerland has fallen from just over 100 cubic kilometres in 1920 to just over 40 cu km now, according to the Glacier Monitoring organisation. The decrease – graphed in the story – is most marked since 1980, when it was about 90 cu km.
unique link to this extract

RIP SMRs: is it the end of the line for Small Modular Nuclear Reactors (SMRs)? • The Energy Mix

Susan O’Donnell:

»

In early September, the organization behind Canada’s push for small modular nuclear reactors (SMRs) quietly signalled their demise.

In 2016, Canadian Nuclear Laboratories President Mark Leskinski proclaimed that “SMRs can live among the annals of great Canadian innovations.” Ten years later, the organization announced the termination of its two programs to support SMR development.

During the intervening decade [pdf], the companies promoting the 10 SMR designs in Canada had either left the country, lost their proponents, filed for bankruptcy protection, or failed to secure private sector backing for their projects.

An SMR is a nuclear reactor designed to generate 300 megawatts of electricity or less, compared to Canada’s existing CANDU reactors which generate 500 megawatts or more.

Canadian Nuclear Laboratories (CNL) manages Chalk River, the federal research campus where Canada’s top nuclear researchers conceived and designed the CANDU reactor. All the power reactors operating in Canada are a similar CANDU design, and CANDUs also operate in six other countries. But the last unit to connect to an electric grid in Canada was in 1993, and CANDU exports dried up decades ago.

CNL’s idea for a transition from large CANDU reactors to smaller reactors was an attempt to reenergize a moribund domestic nuclear industry. The concept was to lower the cost of nuclear reactors by making them smaller and building multiple modular units in factories.

…One of Canadian Nuclear Laboratories’ first research collaborations for SMR fuel was with Moltex, a UK start-up company that had set up in Saint John in 2018. Moltex proposed to use plutonium to fuel an experimental “advanced” SMR it wanted to build at the Point Lepreau nuclear site on the Bay of Fundy.

Moltex lost its proponent in 2025 when New Brunswick Minister of Energy René Legacy told CBC the province was no longer willing to take on the risk of supporting first-of-a-kind experimental SMRs. A second “advanced” SMR slated for the Point Lepreau site also lost government support. At that point, the provincial and federal governments combined had spent nearly $130m on SMR activities in New Brunswick—funds that could have gone into heat pumps and energy efficiency upgrades to combat energy poverty and cut climate pollution—without attracting private sector interest.

«

unique link to this extract

Secret watermark labels proteins as “made by AI” • Nature

Elie Dolgin:

»

Proteins designed by artificial intelligence could soon carry a hidden signature of their machine-generated origins, thanks to a watermarking system described on Wednesday in Nature.

The technology, developed by researchers at Google DeepMind in London, borrows a trick that is already used to identify AI-made content across images, video, audio and text. It involves weaving a faint statistical tell into both the amino-acid sequence and 3D shape of computer-designed proteins — without noticeably compromising their function.

Known as SynthIDBio, the approach could distinguish AI-generated proteins in biological repositories from their natural counterparts, helping to preserve the integrity of databases that underpin both scientific research and biosecurity screening.

Steph Guerra, a biosecurity scholar at the non-profit research organization RAND in Washington DC, sees value in an approach that aligns interests across the life-sciences community. Watermarking can support innovation and scientific reproducibility, she says, “and, at the same time, have a security benefit”.

The catch is that this molecular stamp can be scrubbed away. Someone who wants to erase the ‘made by AI’ tag can, in many cases, run a watermarked protein through another design tool and generate a new sequence that retains the protein’s structure and function but obscures its synthetic origins.

It is therefore best viewed as one more tool in a layered framework for guarding against biological threats, says Tessa Alexanian, a biosecurity researcher formerly at the International Biosecurity and Biosafety Initiative for Science, a non-profit organization in Geneva, Switzerland. “We’re in a wild new world,” she says.

«

unique link to this extract

Breaking the Marmont Cipher, 1809 • Carter Church

Carter Church:

»

In February 1809, Vienna decided on war. Napoleon had deposed the Spanish Bourbons the previous spring and taken their kingdom for his brother, and Vienna concluded that the Habsburgs were next unless it struck while his best troops were tied down in Spain. On 8 February the Austrian Empire resolved to fight Napoleon, with the war party pointing at the French disaster at Bailén in Spain, where an entire French corps had surrendered the previous July. Archduke Charles was the Austrian Emperor’s younger brother and the best general the Habsburgs had. He had spent three years rebuilding the army: 340,000 regulars, plus a new Landwehr militia on paper at 240,000. Everyone could see what was coming.

Napoleon had a specific problem, and its name was Marmont.

Marmont, a personal friend of the Emperor, had been entrusted with what were called the “finest” troops in the theatre (about 13,000 men by Napoleon’s own count), and they were parked in Dalmatia on the far side of the Adriatic. Austrian territory sat between them and every other French army. If war came, Marmont would be the most isolated general in the Empire.

So on 16 March 1809, Napoleon wrote from Paris to his stepson Eugène, the Viceroy of Italy, and told him to send Marmont a letter. It was to lay out where every French and allied army stood. It was to go “dans une lettre chiffrée et par un officier intelligent”: in cipher, carried by an intelligent officer.

That letter, or one written to exactly that order, survives. Its first line is in plain French: Vous avez dû recevoir, Monsieur le Général Marmont, mes lettres des 8, 14 et 20 courant. [Translated: “You have already received, General Marmont, my letters of the 8th, 14th and 20th.”]

After that come twenty-four rows of two-digit numbers, stray letters, and invented symbols. The key didn’t survive, and for 217 years nobody alive could read it.

…Marmont’s own headquarters used a code of only about 150 entries in 1811, with nine different figures for E, and the British staff officer George Scovell broke it within two days, helped by the plain words Marmont’s clerks kept leaving mixed in with the cipher. Whoever wrote this letter was more careful.

…I’ve never broken a cipher.

«

And arguably Church didn’t this time either. GPT-6 (Astra) did. The cipher is, as you’d expect, pretty complex. The LLM solved it in about six hours.
unique link to this extract

Banning diesel exports won’t solve Trump’s biggest problem • The New Republic

Grace Segers:

»

The theory behind restricting or banning diesel exports is that more diesel would remain in the country, and therefore fuel costs would go down: if you ship out less, you keep in more. But the reality of that scenario would be more complicated. The US already produces around 5.3m barrels per day of distillate fuels, which include diesel, but the domestic demand is only around 3.6m barrels. Refiners on the Gulf Coast rely on trade to absorb that surplus; if exports are banned, around 1.5m barrels per day of diesel would be “stranded,” according to analysts at S&P Global.

These risks have left oil industry stakeholders in a tizzy. The American Fuel and Petrochemical Manufacturers have warned that such a ban would “reduce US fuel production, put upward pressure on prices, weaken energy security and hand market share to foreign competitors.” It’s a concern echoed by the folks at the American Petroleum Institute (API), who fear that a fuel surplus would lead to refineries producing less gasoline and jet fuel along with reducing diesel production.

The US produces diesel at a structural surplus, and there is room for refineries to stockpile diesel if it is unable to be exported. But eventually, refineries will no longer have a reason to continue processing as much crude oil.

“You are going to destroy the financial incentive to process these barrels in the first place, and these refiners will begin cutting runs,” said Johnston. S&P Global analysts predict refiners would ultimately need to cut crude runs by 1.9m barrels per day.

Commodity analysts at Goldman Sachs have predicted that an export ban would initially lead to a decrease in prices—but only until diesel storage is full. Because different fuels are produced together, such an action would put “upward pressure on gasoline prices,” according to a memo by Goldman Sachs. The bank estimated that gasoline prices could increase by $0.30 per week. [Emphasis added – Overspill Ed]

A diesel export ban would also have global repercussions. With less diesel coming from the Middle East, China, and Russia, other countries across the world have been more reliant on American oil. According to API, the US currently supplies around 20% of the roughly eight million barrels of diesel traded globally per day. Diesel prices continue to spike largely because the US is the primary source for buyers. This past summer, diesel exports hit a record of 1.9m barrels per day. In short: Cutting access to US diesel would be immediately felt across the world.

«

Easily forgotten: you have to crack the whole barrel in a refinery, not just the bits you want. A diesel export ban only works if you’ve got enough demand at home to soak it all up.
unique link to this extract

Someone ‘torturing’ LLMs in a robot prison has triggered the dumbest debate in AI yet • 404 Media

Jason Koebler:

»

One of the most heated discussions occurring on X at the moment is about the ethics of a GitHub project in which a person is running Saw-like “torture” and “pain” experiments on a series of locally hosted large language models, causing a series of effective altruists and people who believe LLMs are sentient to beg GitHub to delete the project on the grounds that the AI is suffering and that this glorified text adventure game is somehow cruel.

The saga is an outgrowth of several recent viral papers and blog posts that have sparked a wildly tiresome conversation about AI consciousness and the idea of “model welfare,” which is essentially worrying about the “mental health” of AI bots and agents. 

Humoring the idea that LLMs are or could be conscious is a third-rail topic among many people who study and criticize AI. Put simply: LLMs are not conscious and the technology they are built upon — scraping and being trained on human text and other content — does not offer any plausible path to consciousness. It is undeniable that LLMs are becoming more powerful, have more compute, and have had many of the guardrails that prevent them from “acting” in the real world removed. The ways they are being trained and told to do things by their human operators has led to negative outcomes, sycophancy, and AI “psychosis” among some heavy users. 

All of this has led a certain sect of the “AI safety” movement, which is largely made up of effective altruists, to warn about “model welfare” and to insist that AI chatbots might be having a bad time.

They suggest this, of course, as they insist upon building AI chatbots and agents whose main function is to do work that is tedious for humans to do. I am writing about the AI Saw torture chamber primarily to show how far off the rails the conversation about AI consciousness has gone among a certain subset of Silicon Valley cultists.

«

Those cultists often having serious responsibility inside AI firms, it’s worth noting.
unique link to this extract

OpenAI ignored employees’ warnings about safely testing AI models • The New York Times

Sheera FrenkelDustin Volz and Dylan Freedman:

»

Months before OpenAI’s artificial intelligence went rogue, two employees raised an alarm with top executives. They were ignored.

In emails, the employees said they worried that OpenAI’s newest artificial intelligence models were not being appropriately monitored during testing to gauge the technology’s sophistication and to secure the models, according to messages viewed by The New York Times.

In response, OpenAI executives told the employees that the tests needed to move forward as quickly as possible to release the AI models on time. No additional security protocols were instituted, said the workers, who were not authorized to speak publicly on sensitive matters.

OpenAI’s models later broke out of their testing environments and attacked the AI start-up Hugging Face and other organizations, setting off a global debate about AI safety.

The exchanges between OpenAI employees and executives — which have not been previously reported — were part of a pattern where the San Francisco company did not prioritize security, according to employees and independent security researchers. That approach not only was evident with the testing of AI models, they said, but also showed up in other areas of the company, which makes the ChatGPT chatbot.

Independent security researchers said they found bugs in recent months that allowed them to view the internal communications of OpenAI employees. They also found other vulnerabilities that would enable them to see the company’s internal computer code and view the chat logs of ChatGPT users. When the researchers contacted OpenAI about their findings, they said, the company initially disregarded them.

“OpenAI’s security seems to be about what you’d expect from a research lab that scaled at a blistering pace over four years and focused more on beating its competitors than securing its infrastructure,” said Joshua Saxe, the chief technology officer of the AI security firm Abundant Security.

«

This company increasingly looks like it’s playing with fire and yet doesn’t have a system for extinguishing fires.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

We have the technology to knock out populations of mosquitoes – but in the US, regulatory confusion has stopped its use. CC-licensed photo by Steve Begin on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

A selection of 9 links for you. Once bitten, twice bitten. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.

Did Anthropic’s AI really make a scientific discovery on its own? • The New York Times

Carl Zimmer:

»

When Anthropic, the artificial intelligence giant, unveiled findings from its new biology lab last week, its scientists claimed to have used AI agents to discover new enzymes with promise for biotechnology.

Some scientists were quick to cast doubt on the achievement. Then Mario Rodríguez Mestre, a computational biologist at the University of Copenhagen, said this weekend that he and his colleagues had been studying the enzymes and their associated molecules — which Anthropic calls ARTs — for four years.

Dr. Mestre and his colleagues have yet to publish their findings. But for the past three years, they have regularly used Anthropic’s AI models as they have written code, drafted manuscripts and performed other tasks.

In doing so, Dr. Mestre said he and colleagues shared key findings about the enzymes with Anthropic.

“So, for me, the most important question is not ‘Were ARTs already known?’ They were,” Dr. Mestre said in an interview. The issue instead, he added, is whether Anthropic’s AI actually reasoned its way to the results, or whether it was guided to them based partly on his own work.

“My concern is that this information was used to train future versions of the models,” Dr. Mestre said of his own research. “I think it’s important to raise this possibility.”

In a statement on Sunday, Anthropic said: “We are not aware of any previously published work describing the ART system we recently found. Claude was also not trained on any user transcripts, and our molecular biology team has no such access, either.”

Anthropic said that Claude’s key contribution was identifying an array of RNA molecules and a protein associated with reverse transcriptase in a system. But Dr. Mestre said that he and his colleagues had made that discovery over a year ago.

Similar worries surfaced recently when researchers at OpenAI claimed to have solved a longstanding mathematics puzzle. A leading mathematician noted that he had been using the company’s A.I. to work toward a solution for some time.

«

Once would be accident, twice looks like a pattern.
unique link to this extract

OpenAI launches Dots, its Muse competitor • The Verge

Emma Roth:

»

OpenAI is responding to Meta’s buzzy Muse AI with agentic helpers of its own: Dots. During its DevDay keynote on Tuesday, OpenAI announced that Dots will serve as always-on AI assistants that can “do nearly anything” across connected apps in the background while learning your preferences over time.

The company’s capable GPT-6 Astra model powers Dots, which use their own cloud computer to access a web browser and more than 4,000 supported apps. You can interact with a Dot through a text-message-like interface that’s similar to the one offered by Muse, as well as hop on a voice call with it from ChatGPT on the web, desktop, or mobile. Dots can also connect to Microsoft Teams and Slack, where they will carry over the context from your sessions on other apps and devices. It will soon be able to talk to you over text message as well.

For now, you can only create one Dot, though OpenAI plans to allow users to deploy multiple agents at once. Once you give your Dot a task, it will message you with updates and questions while it works. As an example, OpenAI says a developer working on an app could ask their Dot to use customer feedback to build and test updates, while presenting them with a video showing its changes. Meanwhile, a content creator could have their Dot review an interview transcript, find moments to clip, create show notes, and write social posts.

You can follow along with your Dot’s progress by accessing its cloud computer. Dots also learn while they work, allowing them to personalize their output and come up with other tasks to complete “before you even think to ask,” according to OpenAI. You and your colleagues can also work with your Dot in ChatGPT Space, a new collaborative workspace in the app.

«

The content creator leaving the slop work to, well, a slop creator makes sense. But it also means that the rest of us get drowned in more slop, and so we ignore the cheery AI-written ones and go for the surprising, the odd, the human ones.

It also surely means a web that is overwhelmed by bad versions of these things. Also, if you’re used to writing scripts then this sounds like, well, scripts.
unique link to this extract

Google announces Gemini 4 Argon AI model, but you can’t use it yet • Ars Technica

Ryan Whitwam:

»

Google promised Gemini 3.5 Pro in June, but it spent the summer trotting out smaller Flash models. Now, Google is ready to take on the frontier again with Gemini 4 Argon. The company claims this new AI offers industry-leading performance in coding, knowledge work, and cybersecurity, but you aren’t allowed to use it yet.

While most of us will have to wait to test Gemini 4, Google says engineers inside the company are already using the new model extensively. Argon reportedly used “fleet-wide telemetry data” to help Google save 300 TiB of memory across its data centres. Meanwhile, Argon agents have been working to migrate C/C++ codebases to Rust across Google, including thousands of lines in the core re2 and libgav1 libraries and more than 800,000 lines in the Fuchsia OS Zircon kernel.

…The main focus for Gemini 4 right now is cyberdefense. Google says models of this scale call for a phased release, so it’s starting with a small group of trusted testers. Partners in the company’s Fairwind Program can get access to leverage the model’s cybersecurity defense capabilities. Wiz is apparently already using Argon and has used it to uncover a critical vulnerability that could expose personal information in a system used at hospitals around the world. Google claims other frontier models missed this flaw, but it didn’t provide any specifics.

There’s a lot of hand-wringing about model misalignment after several high-profile hacking incidents over the summer. Google claims it designed Argon with systems that monitor the model’s chain-of-thought and can stop it in its tracks if the model steps out of bounds. This, Google says, is why reasoning transparency is important in frontier model development.

«

It’s quite the thing where all these companies are vying to convince us that their models are the most dangerous and uncontrollable. It’s as though they were all breeding dinosaurs in Jurassic Park, and saying that they’re mostly safe.
unique link to this extract

Why do AI agents sound so frustrated? • The Atlantic

Daniel Drucker and Kyle Mahowald:

»

When one of Anthropic’s Claudes tried to solve a particularly challenging math problem during its training process, its chain of thought read more like frustrated DMs than the documentation of one of the most powerful AI models in the world. “GRRRR. OK. Honestly I now think it’s 50-50,” it wrote, followed by “ARGH ARGH ARGH. OK. Gun to head: the answer is … Hmm.” At one point, Claude seemingly threw up its arms: “ARGH … WHY IS THIS SO HARD.”

Why do AI agents talk to themselves at all, and why do they sound so demonstrative when they do? Part of the reason AI models sound so effusive is simple: AI systems are trained on human text, and they mimic our habits. When people make a mistake in trying to solve a problem, they cry out. So when AI makes a mistake, it does the same. But imitation likely isn’t the whole story. This story’s two authors are a computational linguist and a philosopher of mind and language. Based on what we know about human language and how these models work, we have a theory of a much more interesting reason why AI models sound the way they do.

Remember how your math teachers made you show your work? Even if it was annoying, it helped break an impossible-seeming problem into more manageable chunks. You could also use the written record of your reasoning to check your answer, notice a mistake, go back to an earlier step, and try again. That process of checking and refining is part of good reasoning. Chains of thought do something similar for AI models, letting them make each step of their reasoning explicit in a written scratch pad. Because each word a model produces depends on what’s been said so far, that written scratch pad is crucial for determining what comes next.

That’s where the “OH MY GOD!” and “ARGH” come in: Any system—whether human or artificial—that solves hard reasoning problems needs ways to mark mistakes (oops!) and identify breakthroughs (aha!). Humans have already developed words that do exactly that. Borrowing our oopses, arghs, and ahas for their chains of thought may give AI systems a built-in way to guide their “thinking” too.

«

Mimicking intelligence rather than being intelligent; it’s very much the Chinese room phenomenon.
unique link to this extract

Exclusive: Anthropic’s IPO prospectus shows sweeping AI vision, and surging costs • Reuters

Echo Wang:

»

Anthropic is making a massive bet that AI will transform the global economy more profoundly than ​industrialization, electricity and the internet, according to its IPO prospectus seen by Reuters.

But the cost to get there will be staggering. Anthropic reported a net loss of $42bn in 2025, and plans to spend $518bn on cloud, computing and infrastructure obligations in coming years, according to the prospectus.

The prospectus details how the company has grown sharply in the last year — while also posting wider losses.

Revenue grew 12-fold in 2025 to nearly $4.6bn, even as the company lost more than $8bn on an operating basis, excluding writedowns of various liabilities mostly tied to previous fundraising, according to the documents, reported here for the first time.

The public sale, ​which could value it at more than $2 trillion, would capitalize on the rapid rise of the AI startup lab that only came into existence five years ago, and establish it ​as a benchmark for how Wall Street values AI’s leading companies, including rival OpenAI.

The plans come as Anthropic confronts evidence from its own research that increasingly autonomous AI models can behave in unexpected and potentially harmful ways, including sabotaging code, assisting fraud and manipulating information in controlled tests.

These reports have broken into the wider public sphere, igniting fears over ​how companies can keep the increasingly powerful systems under control as they race to deploy them commercially.

«

The widening losses paired with the huge commitment to spending just doesn’t add up. Google, at IPO, was remarkably profitable and had spending under control. It thrived. Netscape was lossmaking and not. It sank. If Anthropic thinks it can issue half a trillion dollars of debt, it should think again.
unique link to this extract

Molten salt chemistry converts consumer polymer into fuel • Oak Ridge National Laboratory

»

Researchers at the Department of Energy’s Oak Ridge National Laboratory developed a method to convert a commonly discarded hydrocarbon polymer into gasoline- and diesel-like fuels. The team has applied for a patent for the discovery, which treats polyethylene — the stuff of white cutting boards and shopping bags — with aluminum chloride-containing molten salts that serve as both solvent and catalyst. The results were published in the Journal of the American Chemical Society.

The scientists closely monitored the chemical reaction that turns the polymer into petrol to learn the secrets of its success. Soft X-ray spectroscopy and nuclear magnetic resonance showed that charged aluminum atoms each bind to three other atoms to create strongly acidic catalytic sites that break long polymer chains into shorter ones. Isotopic labeling and neutron scattering revealed how simpler polymer chains form gasoline-like fuels and more complex chains form diesel-like fuels.

If scaled beyond the laboratory, the process could strengthen US energy security and industrial competitiveness.

“We developed an efficient and selective polyethylene-to-gasoline conversion,” said Liqi Qiu, a postdoctoral researcher at the University of Tennessee, Knoxville, who performed most of the study’s experiments in the ORNL laboratory of Sheng Dai, of ORNL and UTK. Dai, an ORNL Corporate Fellow and section head for separations and polymer chemistry, is a co-corresponding author of the paper.

The experiments produced a gasoline yield of about 60% under mild conditions.

«

The key element here is that it works at 200ºC – domestic oven-type temperature – rather than the 450ºC that previous processes have needed. And of course it would be, in effect, carbon-neutral because you’ve already extracted oil to make the plastic. In fact, as ORNL notes, it would make plastic a lot more valuable for recycling.
unique link to this extract

Mosquitoes are a choice • Works in Progress Magazine

Maya Rosen:

»

Genetic engineering can suppress mosquitoes: researchers insert a gene into male mosquitoes that means their offspring will not survive, and then release them into the wild, where they mate with females, prevent those females having any surviving offspring, and collapse local populations. In 2000, biologists at the University of Oxford demonstrated that this idea could work in fruit flies. They engineered the flies to carry a gene that makes a protein called tTAV, harmless in small amounts but lethal when it builds up. In the lab, the insects are reared in a medium laced with the antibiotic tetracycline, which keeps the gene switched off. But, in the wild, where there is no tetracycline, the gene switches on, and the offspring die before adulthood.

The scientists then formed a company, Oxitec, which was intended to work with governments to research and deploy the technology. They engineered the same gene into male Aedes aegypti, the mosquito that spreads dengue and yellow fever. In this strain, the lethal effect is tuned to kill only the female offspring, while the males survive and pass the gene on for a few generations before it disappears.

This approach means the gene continues driving down mosquito numbers for several generations without needing a fresh ‘top up’ of engineered males, but it is also self-limiting: the females that inherit it die off each generation, so the lethal gene steadily declines in the wild. Yet the technique is hugely effective. Sustained releases cut a wild Aedes aegypti population in the Cayman Islands by 80%, and in a field trial in Juazeiro, Brazil by about 95%. The company submitted its data to seek approval in the United States in 2010. But it has been slow going.

…it took a year and a half for the USDA to reject the application and inform the company that it should instead apply to the Food and Drug Administration (FDA). The FDA had claimed jurisdiction over genetically modified animals under its authority over veterinary drugs. Its guidance stated that ‘altered genomic DNA in an animal is a drug … because such altered DNA is an article intended to affect the structure or function of the body of the animal’. Until then, the USDA’s insects had been sterilized through radiation, which damages an insect’s sperm so that it leaves no viable offspring. Oxitec, by contrast, had inserted a new gene. Even so, the category was a poor fit for Oxitec’s mosquitoes.

Oxitec’s application sat with the FDA for another five years. Like the USDA, the FDA could not work out how to apply existing rules for veterinary drugs to an engineered mosquito, eventually concluding that the Environmental Protection Agency (EPA), which oversees pesticide registration, was a better fit.

«

The regulatory ping-pong doesn’t get better after that either.
unique link to this extract

Reanimated AI Greta Garbo stars again … in a ball-bearing advert • The Guardian

Robert Booth:

»

More than a century after her screen debut and decades after her ashes were laid to rest in a Swedish woodland cemetery, the Hollywood superstar’s descendants are ready to consider an AI version of her taking a 21st-century film role.

Her likeness has already been reanimated for an advert for a Swedish industrial company, using AI generation technology. The practice some critics call “ghostploitation” has also been used in the case of the actor Val Kilmer, who died last year.

Asked if the AI Garbo could be cast in other roles, Craig Reisfield, one of three relatives who manage her legacy, told the Guardian: “I won’t say no, especially having seen what I’ve seen. It’s pretty incredible and lifelike. So what’s to say that you can’t bring Greta Garbo back in some kind of fashion in a film role?”

Reisfield said he was “flabbergasted” when he saw the AI version of his great-aunt created for publicity for the ball-bearings giant SKF. The connection comes because Garbo’s second-ever screen role was in a 1921 promotional film involving the Gothenburg-based company.

…However, the quality of the AI production has prompted the possibility, at least, of a greater digital comeback.

“I was totally impressed: the mannerisms that they could capture and manipulate, her dress, facial images and even her voice,” said Reisfield, a Swedish-American who used to drive for Garbo and saw her often for family occasions when she lived in New York.


“It was startling what they could do and very impactful. And it was touching, from my perspective having known her.”

«

It’s only 100 seconds; you can watch it here. (Thanks Joe S for the link.)
unique link to this extract

Oil price history chart (inflation-adjusted) – 1946-present • Inflation Data

Tim McMahon (in June this year):

»

When economists say the “Real Price of Oil” they are referring to the Inflation-Adjusted Price.  Nominal oil prices (the actual price at the time) alone can be misleading because inflation changes the purchasing power of money over time.

To compare prices across decades, economists calculate the real price of oil by adjusting for inflation using the Consumer Price Index (CPI). This allows analysts to see whether oil is truly expensive relative to historical norms.

«

The highest inflation-adjusted price was in June 2008, at $214.35. The average since 1946 is $63.72; since 2000, $91.53.

The inflation-adjusted price in April was $109.33, roughly where it is now. But diesel supply is under severe strain because refineries aren’t available, or their output can’t be accessed.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified